Filament Monitor Beta Overview
-
We have been working on a filament monitor design that senses filament movement accurately enough to detect a filament jam, filament ground away by the extruder or filament out and pause the print before it's ruined.
The documentation is here:
https://duet3d.com/wiki/Duet3d_Filament_Monitor
The beta enclosure and bigbox mount file for printing are shared here:
https://www.youmagine.com/designs/duet3d-filament-monitorBeta Test
We are sending out beta boards to beta testers and will have some left over, I will start another thread about how to get involved in the beta test.
-
The calibration section on the wiki page is out of date because I forgot to update it. The recommended procedure is now:
- Put the firmware in calibration mode by using a negative R parameter in the M591 command
- Start a print
- During and after the print, M591 D# (where # is the extruder number) will report the measured mm/rev and its variation.
-
Awesome, thats much easier. I have updated the wiki:
https://duet3d.com/wiki/Duet3d_Filament_Monitor#CalibrationI will test the calibration when i get back home.
-
I forgot to mention: any time you pause the print or start a new one, the calibration figures will be reset and restarted when you resume. Otherwise the figures just go on accumulating. At least, that's what I think I implemented. The idea is that if you need to adjust the extruder tension etc. during the calibration, you don't need to start another print to see what effect it has, you can just pause and resume.
-
Please note, it is important that during retraction, the filament reverses through the filament monitor. Otherwise it throws out the commanded/measured movement comparison. So the filament monitor should either be mounted very close to the extruder drive inlet and in a straight line so that the stiffness of the filament forces it back through the monitor, or the Bowden tube connecting the two must be captive at both ends. When using a Titan extruder, there is already a piece of Bowden tube held captive in the extruder inlet. I put a screw-on brass Bowden end from RepRap Ltd on the other end and attached it to the filament monitor body.
-
David, how did you attach the PTC fitting to the filament monitor body? This seems like a quick way for me to connect to my delta.
-
I just lashed them together using 3 cable ties. It isn't very satisfactory because there is still some movement between the brass Bowden end and the filament sensor body. Better would be to print a clip that fits around the slot in the brass piece, then that could be screwed to the filament sensor body.
-
Thanks David. I am printing a modified sensor body that has more plastic to insert (and CA glue) a short piece of PTFE in to.
-
I have it on my todo list to change the mount so you can screw two pneumatic couplings into each side so ti can be inserted inline in a bowden tube…. but i have a long todo list right now!
-
Gentlemen, is this still active? I would be happy to test. I did not find any other thread regarding the beta test. I can build my own pcb's but am less clear regarding the custom-looking hobb and mechanics shown in the documentation.
Cheers,
Ray -
Thanks David. I am printing a modified sensor body that has more plastic to insert (and CA glue) a short piece of PTFE in to.
Does CA glue bond PTFE?
-
Thanks David. I am printing a modified sensor body that has more plastic to insert (and CA glue) a short piece of PTFE in to.
Does CA glue bond PTFE?
Yes (but no in fact)
I scar the outside of the tube to make it "grip"
Edit : sorry DC I've clicked "report" instead of "quote" for this answer
my bad -
Beta Test
We are sending out beta boards to beta testers and will have some left over, I will start another thread about how to get involved in the beta test.
wink wink ;D
-
Not that this particularly interests me as I'd need 5 of them and I really don't see how I'd get another another 15 cables through my cable chain. Oh and all my end stops on the main board are populated. But I'd guess this wouldn't work with a mixing hot end where filaments can be used or unused throughout a print depending on the colour of the part/section that is being printed. Am I correct?
-
Not that this particularly interests me as I'd need 5 of them and I really don't see how I'd get another another 15 cables through my cable chain. Oh and all my end stops on the main board are populated. But I'd guess this wouldn't work with a mixing hot end where filaments can be used or unused throughout a print depending on the colour of the part/section that is being printed. Am I correct?
Yes it should work with a mixing hot end, because each extruder drive is tracked separately. You can connect filament sensors to the 5 endstop connectors on the Duet and to the 2 on the CONN_LCD connector, and you can choose which endstop connector is used for each extruder. If you have endstop switches for XYZ then you have only 4 suitable endstop connectors left, leaving you 1 short.
-
Ah OK. And it could cope with me doing the crazy things I do like varying the mixing ratio on every layer? And a mixing ratio that might be as low as 1% for a particular extruder?
I use end stop switches for the X and Y. The Z homing is a switch but uses the Z probe connector because of its in built LED. Another end stop is connected to switches that are wired in series to the axes maxima and another is connected to an emergency stop button. I guess I could move all these to the Duex5, thus freeing them up on the Duet for the filament sensors? If I read correctly I can't use the end stops on the Duex5 for filament sensors.
-
Currently you can't shift the endstops for the axes. But you can connect a filament sensor to your unused Z endstop input.
-
Currently you can't shift the endstops for the axes. But you can connect a filament sensor to your unused Z endstop input.
Varying mixing ratios and using small mixing proportions shouldn't stop the sensor working.
-
OK. As I said before, it's not something that particularly interests me at this time, and the big deal would be trying to get another 15 wires through my cable chain which already carries the wires for the 5 extruder motors as well as all the hot end gubbins, fans, end stop switches, lights etc. Good to know that it's possibly doable though.
-
@David - re CA for teflon, see https://www.youtube.com/watch?v=Q6mU4excEE8. Turn your audio off though, it is annoying.
-
I printed a small piece to lock in to the groove in the brass Bowden termination:
That allowed me to cable-tie it more securely to the filament monitor block:
EDIT: changed photo to a better one
-
I'm building my first Delta printer and am very keen to add filament monitoring so have been following this thread with great interest. I've looked all over for the separate thread about becoming a beta tester but can't find it - are there still boards available and, if so, how do I get involved, please?
-
Why does the sensor only work on the Duet, not the Duex? Is it because you re-configure inputs as analog, which is not possible on Duex?
-
It's because the pins on the DueX have too much latency, because they are read via a I2C I/O expander.
-
I see.
